Type 2 diabetes (T2D) has reached epidemic proportion. Lifestyle factors such as a healthy diet are crucial for T2D prevention. Recently, the PREDIMED (PREvencion con DIeta MEDiterranea) trial was completed in Spain. The trial was a multicenter, randomized controlled trial in which 7,447 participants were randomly allocated to Mediterranean dietary interventions enriched with extra virgin olive oil or nuts or a control diet. Although the trial was primarily designed to assess the effect of the interventions on CVD, the dietary interventions also reduced incidence of T2D by 30% (95% CI 0.54-0.92). The goal of the present study is to examine whether the effects of Mediterranean diet on changes in metabolites and T2D risk. We will employ a case-cohort study design with 273 incident cases of T2D and 708 randomly selected participants free of diabetes at baseline (20%). Stored fasting blood specimens collected at baseline and at year 1 of the trial will be measured for a well-validated panel of ~300 metabolites, using the state-of-the-art LC-MS technology. Our Specific Aims are: 1) To examine the association between the baseline metabolite concentrations and the risk of T2D using a case-cohort design; 2). To examine whether the dietary interventions modify the relationships between baseline levels of metabolites and subsequent T2D risk using a case- cohort design; 3). To examine whether 1-year changes in metabolites mediate the effect of the dietary interventions on subsequent T2D outcomes from years 2 to 5 using a case-cohort design; and 4). To examine whether 1-year changes in metabolites influence insulin resistance and -cell function from years 2 to 5 among 708 randomly selected participants free of diabetes at baseline. The large PREDIMED trial, with archived plasma samples and rich repeated measurements of diet, lifestyle, and biomarkers, represents an unprecedented opportunity to address our specific aims in an extremely cost-effective manner. This project has great potential to advance our understanding of the metabolic pathways through which dietary interventions influence T2D risk. It also has important public health and clinical implications because many of these metabolites and their associated pathways may be suitable for direct diet and lifestyle interventions. To prepare for this project, we have already completed a preliminary study by randomly selecting 20 samples from each diet group at one year of intervention and conducted metabolomic analyses using LC-MS at the Broad Institute of MIT and Harvard.